YEBEH-05437; No of Pages 7

Epilepsy & Behavior xxx (2017) xxx–xxx

Contents lists available at ScienceDirect

Epilepsy & Behavior

journal homepage: www.elsevier.com/locate/yebeh

Comorbidities and risk factors associated with newly diagnosed epilepsy

in the U.S. pediatric population
Ahyuda Oh a, David J. Thurman b, Hyunmi Kim a,b,c,⁎
a
Department of Pediatrics, Emory University School of Medicine, Atlanta, GA, USA
b
Department of Neurology, Emory University School of Medicine, Atlanta, GA, USA
c
Division of Pediatric Neurology, Children's Healthcare of Atlanta, Atlanta, GA, USA

a r t i c l e i n f o a b s t r a c t

Article history: Neurobehavioral comorbidities can be related to underlying etiology of epilepsy, epilepsy itself, and adverse ef-
Received 26 June 2017 fects of antiepileptic drugs. We examined the relationship between neurobehavioral comorbidities and putative
Accepted 19 July 2017 risk factors for epilepsy in children with newly diagnosed epilepsy. We conducted a retrospective analysis of chil-
Available online xxxx
dren aged ≤18 years in 50 states and the District of Columbia, using the Truven Health MarketScan® commercial
claims and encounters database from January 1, 2009 to December 31, 2013. The eligible study cohort was con-
Keywords:
Childhood epilepsy
tinuously enrolled throughout 2013 as well as enrolled for any days during a baseline period of at least the prior
Comorbidity 2 years. Newly diagnosed cases of epilepsy were defined by International Classification of Diseases, Ninth Revision,
Newly diagnosed epilepsy Clinical Modification-coded diagnoses of epilepsy or recurrent seizures and evidence of prescribed antiepileptic
Risk factors drugs during 2013, when neither seizure codes nor seizure medication claims were recorded during baseline
periods. Twelve neurobehavioral comorbidities and eleven putative risk factors for epilepsy were measured.
More than 6 million children were analyzed (male, 51%; mean age, 8.8 years). A total of 7654 children were iden-
tified as having newly diagnosed epilepsy (125 per 100,000, 99% CI = 122–129). Neurobehavioral comorbidities
were more prevalent in children with epilepsy than children without epilepsy (60%, 99% CI = 58.1–61.0 vs.
23%, CI = 23.1–23.2). Children with epilepsy were far more likely to have multiple comorbidities (36%, 99%
CI = 34.3–37.1) than those without epilepsy (8%, 99% CI = 7.45–7.51, P b 0.001). Preexisting putative risk factors
for epilepsy were detected in 28% (99% CI = 26.9–29.6) of children with epilepsy. After controlling for demo-
graphics, neurobehavioral comorbidities, family history of epilepsy, and other risk factors than primary interest,
neonatal seizures had the strongest independent association with the development of epilepsy (OR = 29.8, 99%
CI = 23.7–37.3, P b 0.001). Compared with children with risk factors but no epilepsy, those with both epilepsy
and risk factors were more likely to have intellectual disabilities (OR = 13.4, 99% CI = 11.9–15.0, P b 0.001).
The epilepsy and intellectual disabilities could share the common pathophysiology in the neuronal network.
© 2017 Elsevier Inc. All rights reserved.

1. Introduction neurobehavioral comorbidities and risk factors for epilepsy remain to

Epilepsy is one of the most common chronic neurological disorders,
having a significant public health burden [1–3]. Children are the largest
subpopulation with newly diagnosed epilepsy [4]. Recurrent epileptic
seizures compromise health, development, education, and quality of
life [5–7]. Furthermore, a variety of neurobehavioral comorbidities
may adversely affect life performance and quality of life as much or
more than seizures themselves [8–13]. Despite intellectual disabilities
and behavioral disorders that are commonly known to coexist in chil-
dren with newly diagnosed epilepsy [9,14–16], relationships between
be delineated in children with newly diagnosed epilepsy.
Using a U.S. large nationwide healthcare claims database, we
sought to determine if there is a difference in the occurrence of neuro-
behavioral comorbidities between children with and without risk fac-
tors for epilepsy. We also sought to estimate the conditional risk for
neurobehavioral comorbidities given the presence of epilepsy and risk
factors for epilepsy, while controlling for age, sex, residence, and family
history of epilepsy.
2. Methods

2.1. Study population

⁎ Corresponding author at: Department of Pediatrics, Emory University School of
Medicine, 1605 Chantilly Drive NE, Suite 300, Atlanta, GA 30324, USA. We conducted a retrospective observational study of children aged
E-mail address: hyunmi.kim@emory.edu (H. Kim). 0–18 years enrolled between January 1, 2009 and December 31, 2013

http://dx.doi.org/10.1016/j.yebeh.2017.07.040
1525-5050/© 2017 Elsevier Inc. All rights reserved.

Please cite this article as: Oh A, et al, Comorbidities and risk factors associated with newly diagnosed epilepsy in the U.S. pediatric population,
Epilepsy Behav (2017), http://dx.doi.org/10.1016/j.yebeh.2017.07.040
2 A. Oh et al. / Epilepsy & Behavior xxx (2017) xxx–xxx
in the Truven Health MarketScan® commercial claims and encounters
database, which collects data from across the United States. The data-
base contains information from active employees, early retirees, the
Consolidated Omnibus Budget Reconciliation Act (COBRA) continues,
and their dependents insured by employer-sponsored healthcare
plans. The information captures person-specific enrollment records
with demographics and enrollment status, clinical utilization, and ex-
penditures across inpatient, outpatient, and pharmaceutical dispensing
care venues.
All data are compliant with the Health Insurance Portability and
Accountability Act (HIPAA). This study was approved and Institutional
Review Board (IRB) exemption was obtained by the Emory Institutional
Review Board.
2.2. Inclusion criteria and case identification
We included children 0–18 years of age in 50 states and the District
of Columbia, who were continuously enrolled throughout 2013 as well
as enrolled for any days during at least the prior 2 years. Subjects under
1 year of age needed no prior year of enrollment, while those aged 1 to
b2 years required at least 1 prior year of enrollment (Fig. 1).
Epilepsy cases were defined consistent with the guidelines of the
International League Against Epilepsy (ILAE) Epidemiology Commission
[17] and a computer algorithm developed for detecting epilepsy cases
[18]. Using International Classification of Diseases, Ninth Revision,
Clinical Modification (ICD-9-CM) diagnosis codes and antiepileptic
drug claims, we identified a case of epilepsy if it met any of the following
criteria: 1) an occurrence of ≥2 ICD-9-CM codes 345.xx among separate
medical encounters on separate dates; 2) an occurrence of ≥1 ICD-9-CM
codes 345.xx and ≥1 ICD-9-CM codes 780.39 among separate medical
encounters on separate dates; 3) an occurrence of 1 ICD-9-CM code
345.xx and code(s) for antiepileptic drug prescription; and 4) an occur-
rence of ≥2 ICD-9-CM codes 780.39 among separate medical encounters
Fig. 1. Study flow diagram.
Please cite this article as: Oh A, et al, Comorbidities and risk factors associated with newly diagnosed epilepsy in the U.S. pediatric population,
Epilepsy Behav (2017), http://dx.doi.org/10.1016/j.yebeh.2017.07.040
on separate dates and code(s) for antiepileptic drug prescription [19,20].
Subjects with newly diagnosed epilepsy were defined as those who
satisfied the epilepsy case definition in 2013 but with neither seizure
codes nor seizure medication claims during their baseline periods.
We identified twelve neurobehavioral comorbidities and eleven pu-
tative risk factors for epilepsy from the literature. The neurobehavioral
comorbidities included intellectual disabilities (mental retardation,
specific delay in development, delayed milestones, and problem with
learning), psychiatric & behavioral disorders (hyperkinetic syndrome
of childhood, conduct disorder, behavioral problem, anxiety & emo-
tional disturbances, depression, and autism), and headache & migraine.
A case of comorbidity was defined as a subject having ≥1 corresponding
ICD-9-CM codes of comorbidity in either 2013 or any previous years.
The putative risk factors for epilepsy included short gestation & low
birth weight, traumatic brain injury, congenital anomaly, metabolic
disorder, childhood static encephalopathy (i.e., cerebral palsy), central
nervous system (CNS) infection, Malignant tumor or other brain
tumors, stroke or cerebrovascular disease, perinatal morbidity, and
neonatal seizures. Subjects were determined as having a preexisting
risk factor for epilepsy if ≥1 corresponding ICD-9-CM codes of risk factor
were recorded either before the epilepsy index date or b 90 days after
the epilepsy index date. The epilepsy index date was determined by
the first appearance of ICD-9-CM code 345.xx or the second appearance
of ICD-9-CM code 780.39.
2.3. Other variables of interest
Demographic variables included sex, age, and residence. Age was
categorized into the following four groups: 0–4, 5–9, 10–14, and 15–
18 years of age. Based on the 2010 U.S. Census [21], subjects' residential
areas were divided into three groups, including metropolitan statistical
areas (MSAs) with a population of ≥1 million, MSAs with a population of
b1 million, and non-MSAs. The Truven Health MarketScan® database
 A. Oh et al. / Epilepsy & Behavior xxx (2017) xxx–xxx 3

has a variable of unique family identification, which enables us to iden-
tify subjects in the same household. Using the family identification
variable, we coded family history of epilepsy as “yes” if at least two sub-
jects with the same family identification number had epilepsy.
2.4. Statistical analysis
Descriptive statistics were reported as means with standard devia-
tions for continuous variables and frequencies with percentages for
categorical variables. Subject characteristics were compared between
children with and without newly diagnosed epilepsy, using Chi-Square
Tests. Among children with epilepsy, differences in the occurrence of
neurobehavioral comorbidities by risk factors for epilepsy were evalu-
ated with Chi-Square Tests. Multiple logistic regression analyses were
conducted to measure the conditional risk for comorbidities as well as
impact of preexisting putative risk factors on newly diagnosed epilepsy.
We tested the coefficients in four different regression models using a
combination of covariates, such as demographics, comorbidity, family
history of epilepsy, and other risk factors than primary interest that affect
the significance and magnitude of risk factors for epilepsy. Significance
level was set at α = 0.01 with a 2-tailed test. Statistical analysis was
performed using IBM® SPSS® Statistics 23 (Armonk, NY: IBM Corp.).
3. Results
3.1. Subject characteristics
The study cohort comprised 6,107,678 subjects, 3,122,319 of whom
were male [51%, 99% confidence interval (CI) = 51.1–51.2]. The mean
age was 8.8 years (standard deviation = 5.9 years). More than half of
the subjects resided in metropolitan statistical areas (MSAs) with a pop-
ulation of ≥1 million (52%, 99% CI = 51.6–51.7). A total of 7654 children
were identified as having newly diagnosed epilepsy (125 per 100,000,
99% CI = 122–129).
Neurobehavioral comorbidities were more frequent in children with
epilepsy than children without epilepsy (60%, 99% CI = 58.1–61.0 vs.
23%, CI = 23.1–23.2). Children with epilepsy were far more likely
to have multiple comorbidities (36%, 99% CI = 34.3–37.1) than those
without epilepsy (8%, 99% CI = 7.45–7.51, P b 0.001). Preexisting
putative risk factors for epilepsy were detected only in 28% (99% CI =
26.9–29.6) of children with epilepsy, who were most likely than those
without epilepsy to have two or more risk factors for epilepsy (12%,
99% CI = 10.7–12.6 vs. 1%, 99% CI = 0.97–0.99, P b 0.001) (Table 1).
3.2. Neurobehavioral comorbidity burden in children with epilepsy
To avoid possible multicollinearity, we have categorized twelve co-
morbidities into three groups: intellectual disabilities (mental retarda-
tion, specific delay in development, delayed milestones, and problem
with learning), psychiatric & behavioral disorders (hyperkinetic syn-
drome of childhood, conduct disorder, behavioral problem, anxiety &
emotional disturbances, and depression), and headache & migraine.
Among children with epilepsy, psychiatric & behavioral disorders
were most prevalent (30%, 99% CI = 28.7–31.4), followed by headache
& migraine (27%, 99% CI = 25.9–28.5) and intellectual disabilities (26%,
99% CI = 24.3–26.8). When comparing the occurrence of comorbidities
between children with and without epilepsy, the biggest gap was found
in intellectual disabilities (26% vs. 6%, P b 0.001), followed by headache
& migraine (27% vs. 9%, P b 0.001) and psychiatric & behavioral disorder
(30% vs. 12%, P b 0.001).
Distribution of comorbidities by epilepsy incidence age was depicted
on Fig. 2. In the group of intellectual disabilities, specific delay in devel-
opment and delayed milestones were frequently coded at an early age
and decreased afterwards (Fig. 2A). Problems with learning began to

Table 1
Subject characteristics and comparisons between those with and without newly diagnosed epilepsy.

Characteristics All (n = 6,107,678) Comparisons P-value

Without epilepsy (n = 6,100,024) With newly diagnosed epilepsy (n = 7654)

Sex, no. (%)

Male 3,122,319 (51) 3,118,264 (51) 4055 (53) 0.001
Female 2,985,359 (49) 2,981,760 (49) 3599 (47)
Age (years), mean (SD) 8.8 (5.9) 8.8 (5.9) 8.2 (6.2)
0–4, no. (%) 1,818,384 (30) 1,815,698 (30) 2686 (35) b0.001
5–9 1,358,508 (22) 1,356,908 (22) 1600 (21)
10–14 1,560,325 (26) 1,558,662 (26) 1663 (22)
15–18 1,370,461 (22) 1,368,756 (22) 1705 (22)
Residence, no. (%)
Non-MSAs 773,630 (13) 772,553 (13) 1077 (14) b0.001
MSAs (b1 million)a 2,177,332 (36) 2,174,513 (36) 2819 (37)
MSAs (≥1 million)b 3,156,698 (52) 3,152,940 (52) 3758 (49)
Family history of epilepsy, no. (%)
No 6,104,692 (100) 6,097,221 (100) 7471 (98) b0.001
Yes 2986 (0) 2803 (0) 183 (2)
Number of comorbidityc per person, no. (%)
0 4,688,986 (77) 4,685,888 (77) 3098 (41) b0.001
1 959,635 (16) 957,813 (16) 1822 (24)
2 300,798 (5) 299,430 (5) 1368 (18)
3 104,738 (2) 103,982 (2) 756 (10)
≥4 53,521 (1) 52,911 (1) 610 (8)
Number of risk factorsd per person, no. (%)
0 5,762,981 (94) 5,757,489 (94) 5492 (72) b0.001
1 284,042 (5) 282,774 (5) 1268 (17)
2 49,731 (1) 49,273 (1) 458 (6)
≥3 10,924 (0) 10,488 (0) 436 (6)

Abbreviations: MSAs, metropolitan statistical areas; SD, standard deviation.

a
MSAs (b1 million) refers to MSAs with a population of b1 million.
b
MSAs (≥1 million) refers to MSAs with a population of ≥1 million.
c
Comorbidity refers to neurobehavioral comorbidities, including mental retardation, specific delay in development, delayed milestones, problem with learning, hyperkinetic syndrome
of childhood, conduct disorder, behavioral problem, anxiety & emotional disturbances, depression, autism, headache, and migraine.
d
Risk factors refer to preexisting risk factors for epilepsy, such as short gestation & low birth weight, traumatic brain injury, congenital anomaly, metabolic disorder, childhood static
encephalopathy, CNS infection, malignant brain tumor, other brain tumors, stroke or cerebrovascular disease, perinatal morbidity, and neonatal seizures.

Please cite this article as: Oh A, et al, Comorbidities and risk factors associated with newly diagnosed epilepsy in the U.S. pediatric population,
Epilepsy Behav (2017), http://dx.doi.org/10.1016/j.yebeh.2017.07.040
4 A. Oh et al. / Epilepsy & Behavior xxx (2017) xxx–xxx

Fig. 2. Distribution of comorbidities by epilepsy incidence age. A, Intellectual disabilities. B, Psychiatric & behavioral disorders. C, Headache & migraine.

Please cite this article as: Oh A, et al, Comorbidities and risk factors associated with newly diagnosed epilepsy in the U.S. pediatric population,
Epilepsy Behav (2017), http://dx.doi.org/10.1016/j.yebeh.2017.07.040
 A. Oh et al. / Epilepsy & Behavior xxx (2017) xxx–xxx 5

Table 2 (OR = 2.7, 99% CI = 1.9–3.9, P b 0.001), traumatic brain injury (OR =
Neurobehavioral comorbidities in children with newly diagnosed epilepsy, by preexisting 1.6, 99% CI = 1.4–1.9, P b 0.001), and perinatal morbidity (OR = 1.2,
risk factors for epilepsy.
99% CI = 1.0–1.5, P b 0.01). On the other hand, subjects with short ges-
Comorbidity, no. (%) With risk Without risk P-value tation & low birth weight were less likely to develop epilepsy (OR = 0.8,
factors factors 99% CI = 0.66–0.98, P b 0.01).
(n = 2162) (n = 5492)

Intellectual disabilities 947 (44) 1008 (18) b0.001 3.4. Comorbidities according to risk factors
Specific delay in development 681 (32) 670 (12) b0.001
Delayed milestones 605 (28) 430 (8) b0.001
Mental retardation 216 (10) 209 (4) b0.001 Among children with epilepsy, intellectual disabilities more fre-
Problem with learning 59 (3) 142 (3) 0.724 quently occurred in those having risk factors for epilepsy than those
Psychiatric & behavioral disorders 545 (25) 1755 (32) b0.001 with not (P b 0.001). We found the same results in specific delay in
Hyperkinetic syndrome of childhood 273 (13) 973 (18) b0.001 development, delayed milestones, and mental retardation (Table 2).
Anxiety & emotional disturbances 246 (11) 776 (14) 0.001
Depression 225 (10) 671 (12) 0.027
On the contrary, children without risk factors were more likely to
Conduct disorder 112 (5) 313 (6) 0.372 have psychiatric & behavioral disorders, specifically hyperkinetic syn-
Autism 74 (3) 237 (4) 0.075 drome of childhood (P b 0.001) and anxiety/emotional disturbances
Behavioral problem 23 (1) 77 (1) 0.241 (P = 0.001). Headache & migraine presented regardless of risk factors
Headache & migraine 577 (27) 1504 (27) 0.537
for epilepsy.
Headache 548 (25) 1382 (25) 0.868
Migraine 178 (8) 550 (10) 0.017 Neurobehavioral comorbidities were more prevalent in children
having both epilepsy and putative risk factors than in those having
neither epilepsy nor the risk factors, while controlling for sex, age, resi-
be coded at preschool age. Mental retardation was more frequently re- dence, and family history of epilepsy (Table 3). Compared with children
corded at b1 year of age, decreased until 2 years of age, and remained with risk factors but no epilepsy, children with both epilepsy and risk
at consistent rates afterwards. In the group of psychiatric & behavioral factors were more likely to have intellectual disabilities (OR = 13.4,
disorders, hyperkinetic syndrome appeared to be diagnosed around 2 99% CI = 11.9–15.0, P b 0.001).
to 3 years of age and then increased markedly up to 9 years of age
(Fig. 2B). Depression and anxiety disorders increased gradually with 4. Discussion
age. There were no big changes in the occurrence of conduct disorder,
behavioral problem, and autism over age. The prevalence of headache In this large cohort, neurobehavioral comorbidities were significant-
& migraine continuously increased with age (Fig. 2C). ly more prevalent in children with newly diagnosed epilepsy, 2.5 times
the rate of children without epilepsy. Children with epilepsy were
3.3. Impact of preexisting risk factors on newly diagnosed epilepsy 4.5 times more likely to have ≥ 2 comorbidities compared with those
without epilepsy.
After controlling for sex, age, residence, neurobehavioral comorbidity We observed that the occurrence of neurobehavioral comorbidities
groups, family history of epilepsy, and other risk factors than primary in- in children with epilepsy varied across ages and comorbidity groups.
terest, neonatal seizures had the strongest independent association with In the case of intellectual disabilities, specific delay in development
developing epilepsy in this cohort (odds ratio, OR = 29.8, 99% CI = 23.7– and delayed milestones were most commonly detected among children
37.3, P b 0.001), followed by CNS infection (OR = 6.6, 99% CI = 5.2–8.5, with epilepsy onset 6 years of age or less, whereas a problem with learn-
P b 0.001) and childhood static encephalopathy (OR = 5.4, 99% CI = ing was more commonly diagnosed among older children. We found an
4.6–6.3, P b 0.001). The risk of developing epilepsy was also increased opposite pattern in psychiatric & behavioral disorders. Hyperkinetic
in malignant brain tumor (OR = 3.4, 99% CI = 2.3–5.0, P b 0.001), stroke syndrome of childhood appeared to be diagnosed around 3 to 4 years
or cerebrovascular disease (OR = 3.1, 99% CI = 2.6–3.6, P b 0.001), met- of age and increased remarkably up to 9 years of age. Anxiety disorders
abolic disorder (OR = 3.1, 99% CI = 2.4–4.1, P b 0.001), congenital and depression increased gradually with ages and rose sharply at early
anomaly (OR = 2.8, 99% CI = 2.4–3.2, P b 0.001), other brain tumors teen (13–14 years of age). In general, conduct disorder, autism, and

Table 3
Conditional risks for neurobehavioral comorbidities.

Comorbidity, adjusted OR [99% CI]a Epilepsy (+) Epilepsy (+) Epilepsy (−)

Risk factors for epilepsy (+) Risk factors for epilepsy (−) Risk factors for epilepsy (+)
(n = 2162) (n = 5492) (n = 342,535)

Intellectual disabilities 13.4 [11.91–15.01]⁎⁎⁎ 4.3 [3.90–4.69]⁎⁎⁎ 2.8 [2.79–2.86]⁎⁎⁎

Delayed milestones 33.6 [29.52–38.24]⁎⁎⁎ 8.3 [7.23–9.42]⁎⁎⁎ 5.6 [5.47–5.71]⁎⁎⁎
Specific delay in development 18.0 [15.89–20.49]⁎⁎⁎ 5.9 [5.26–6.56]⁎⁎⁎ 3.5 [3.43–3.56]⁎⁎⁎
Problem with learning 7.1 [5.02–10.02]⁎⁎⁎ 4.8 [3.85–5.99]⁎⁎⁎ 2.3 [2.14–2.36]⁎⁎⁎
Mental retardation 4.3 [3.56–5.21]⁎⁎⁎ 2.0 [1.66–2.41]⁎⁎⁎ 1.7 [1.69–1.77]⁎⁎⁎
Psychiatric & behavioral disorders 4.1 [3.56–4.76]⁎⁎⁎ 3.9 [3.59–4.23]⁎⁎⁎ 1.7 [1.67–1.72]⁎⁎⁎
Autism 7.7 [5.66–10.54]⁎⁎⁎ 7.9 [6.61–9.38]⁎⁎⁎ 2.1 [2.02–2.22]⁎⁎⁎
Conduct disorder 4.8 [3.76–6.24]⁎⁎⁎ 4.3 [3.70–5.02]⁎⁎⁎ 1.8 [1.73–1.84]⁎⁎⁎
Depression 4.6 [3.78–5.65]⁎⁎⁎ 3.9 [3.48–4.37]⁎⁎⁎ 1.7 [1.66–1.73]⁎⁎⁎
Anxiety & emotional disturbances 4.5 [3.77–5.47]⁎⁎⁎ 4.1 [3.68–4.53]⁎⁎⁎ 1.9 [1.82–1.89]⁎⁎⁎
Hyperkinetic syndrome of childhood 3.1 [2.59–3.72]⁎⁎⁎ 3.2 [2.93–3.56]⁎⁎⁎ 1.6 [1.52–1.58]⁎⁎⁎
Behavioral problem 2.6 [1.50–4.44]⁎⁎⁎ 2.9 [2.17–3.93]⁎⁎⁎ 1.7 [1.59–1.78]⁎⁎⁎
Headache & migraine 7.0 [6.04–8.06]⁎⁎⁎ 4.5 [4.17–4.93]⁎⁎⁎ 3.7 [3.61–3.70]⁎⁎⁎
Migraine 7.5 [6.02–9.27]⁎⁎⁎ 6.5 [5.72–7.30]⁎⁎⁎ 2.8 [2.70–2.84]⁎⁎⁎
Headache 7.0 [6.04–8.07]⁎⁎⁎ 4.4 [4.03–4.79]⁎⁎⁎ 3.8 [3.71–3.82]⁎⁎⁎

Abbreviations: OR, odds ratio; CI, confidence interval.

a
Reference group was subjects having neither epilepsy nor preexisting risk factors for epilepsy (n = 5,757,489); adjusted odds ratio was calculated, while controlling for sex, age,
residence, and family history of epilepsy.
⁎⁎⁎ P b 0.001.

Please cite this article as: Oh A, et al, Comorbidities and risk factors associated with newly diagnosed epilepsy in the U.S. pediatric population,
Epilepsy Behav (2017), http://dx.doi.org/10.1016/j.yebeh.2017.07.040
6 A. Oh et al. / Epilepsy & Behavior xxx (2017) xxx–xxx
behavioral problem were uniformly distributed across ages. Headache
and migraine continuously increased over ages. This is consistent with
findings from a recent population-based study reporting an age distri-
bution of comorbidities in childhood epilepsy, where psychiatric disor-
ders and headache increased by age [22]. Another recent case–control
study also showed a higher proportion of anxiety and depression in
older children (aged 13–18 years) with epilepsy as well as a lower fre-
quency of cognitive dysfunction and intellectual disability compared
with younger children (aged 6–12 years) with epilepsy [23]. Our finding
might reflect a coding pattern of comorbid conditions, presenting a
possibility of overlapping or duplicating of disease coding on the same
condition.
Children with neonatal seizures were found to be at the highest risk
of developing epilepsy in this cohort. Our findings are consistent with
those of another population-based live birth cohort study, which also
found that some factors happening during pre- or neonatal periods
were highly associated with the occurrence of childhood epilepsy, and
one of the most significant risk factors was neonatal seizures [24].
Healthcare during pre-, peri-, and neonatal periods appear to be impor-
tant to prevent neonatal seizures and epilepsy later in life. In this cohort,
the majority (72%) of children with epilepsy had no preexisting putative
risk factors. Plausible explanations might be the intrinsic limitation of
claims data (e.g., coding errors), short study periods, and other putative
risk factors not included in this study. Further study on revealing the un-
known risk factors for epilepsy is warranted.
While conducting multiple logistic regression analyses in order
to measure the impact of risk factors on epilepsy, we controlled not
only for sex, age, residence, neurobehavioral comorbidities, and family
history of epilepsy but also for other risk factors than primary interest.
This appeared necessary, considering that 41% of children with epilepsy
and any risk factor were found to have more than 2 risk factors, which
may confound each other. We, thus, tried to cancel out the influence
of multiple risk factors per person and then examine the effect of each
primary interest of risk factor, all other observed factors being equal.
This might have an impact on the magnitude of short gestation and
low birth weight coefficient. The prior studies also showed that short
gestation & low birth weight were not a risk factor for developing
epilepsy among children with cerebral palsy [25,26]. These findings
along with ours suggest that it is important to look at the risk factor
while controlling other confounding risk factors.
In children with epilepsy, neurobehavioral comorbidities can
be related to underlying etiology of epilepsy, seizures themselves
(i.e., hyperexcitable networks of neurons), and adverse effects of antiep-
ileptic drugs. Because this cohort included newly diagnosed epilepsy
cases, antiepileptic drugs of short duration are less likely influential.
Among epilepsy cases, the occurrence of comorbidities was compared
between those with and without preexisting putative risk factors
for epilepsy. Intellectual disabilities were much more common in
those having risk factors for epilepsy. This might indicate that intellec-
tual disabilities—especially specific delay in development, delayed
milestones, and mental retardation—are associated with an underlying
etiology of epilepsy rather than epilepsy itself or adverse effects of
antiepileptic drugs. On the other hand, psychiatric & behavioral
disorders—specifically hyperkinetic syndrome of childhood and anxiety/
emotional disturbances—more frequently occurred in those without
epilepsy risk factors. This suggests that psychiatric & behavioral disor-
ders might substantially relate to seizure occurrence per se, not just
the underlying etiology of epilepsy. However, this finding can also be
explained by the possibility that children with a relatively low cogni-
tive function might be underdiagnosed for anxiety/emotional distur-
bances and depression. Headache and migraine presented regardless
of epilepsy risk factors, which suggests its association with epilepsy
itself, underlying etiology of epilepsy, or adverse effects of antiepileptic
drugs.
Compared with children with epilepsy risk factors but no epilepsy,
children with both epilepsy and its risk factors were more likely to
Please cite this article as: Oh A, et al, Comorbidities and risk factors associated with newly diagnosed epilepsy in the U.S. pediatric population,
Epilepsy Behav (2017), http://dx.doi.org/10.1016/j.yebeh.2017.07.040
have neurobehavioral comorbidities in this cohort. The adjusted odds
ratio was nearly 5 times higher in intellectual disability while about
2 times higher in psychiatric/behavioral disorders and headache/
migraine. This suggests that epilepsy and neurobehavioral comorbidity
may share the common pathophysiology through hyper-excitable net-
work of neurons.
The current study has the following limitations. First, Medicaid
data were not included in this study. The proportion of children (aged
0–18 years) covered by employer-based health insurance and Medicaid
in the United State is estimated by 49% and 38%, respectively [27].
Second, there are intrinsic shortfalls of claims data, such as miscoding
and underreporting epilepsy. Third, the study periods were short:
a baseline assessment of up to 4 years of data preceding the epilepsy in-
cidence year may not suffice, resulting in some prevalent epilepsy cases
misclassified as newly diagnosed epilepsy cases. This also can influence
the identification of comorbidities in new cases of epilepsy. Considering
the latency between neurological insults and seizures, identification of
preexisting risk factors for epilepsy might be underestimated.
5. Conclusions
Our study confirms a high prevalence of neurobehavioral comor-
bidities in children with newly developed epilepsy. As we found that
intellectual disabilities were much more common in those having risk
factors for epilepsy, the epilepsy and intellectual disability could share
the common pathophysiology in the neuronal network. By recognizing
and understanding comorbidities, parents or caregivers of children with
newly diagnosed epilepsy can prepare their children and themselves for
changes in treatment goals and the need of additional care.
Funding
This work was supported by a Children's Healthcare of Atlanta
Research Grant from the Goizueta Foundation (Intramural grant ID #
640052-0816).
Conflict of interest
The authors have no potential conflict of interest relevant to this ar-
ticle to disclose.
References
[1] Austin JK, Carr DA, Hermann BP. Living well II: a review of progress since 2003.
Epilepsy Behav 2006;9:386–93.
[2] England MJ, Austin JK, Beck V, Escoffery C, Hesdorffer DC. Erasing epilepsy stigma:
eight key messages. Health Promot Pract 2014;15:313–8.
[3] Koh HK, Kobau R, Whittemore VH, Mann MY, Johnson JG, Hutter JD, et al. Toward an
integrated public health approach for epilepsy in the 21st century. Prev Chronic Dis
2014;11:E146.
[4] IOM (Institute of Medicine). Epilepsy across the spectrum: promoting health and
understanding. Washington, DC: The National Academies Press; 2012.
[5] Kerr C, Nixon A, Angalakuditi M. The impact of epilepsy on children and adult
patients' lives: development of a conceptual model from qualitative literature.
Seizure 2011;20:764–74.
[6] Cianchetti C, Messina P, Pupillo E, Crichiutti G, Baglietto MG, Veggiotti P, et al. The
perceived burden of epilepsy: impact on the quality of life of children and adoles-
cents and their families. Seizure 2015;24:93–101.
[7] Payne ET, Zhao XY, Frndova H, McBain K, Sharma R, Hutchison JS, et al. Seizure bur-
den is independently associated with short term outcome in critically ill children.
Brain 2014;137:1429–38.
[8] Hermann B, Seidenberg M, Jones J. The neurobehavioural comorbidities of epilepsy:
can a natural history be developed? Lancet Neurol 2008;7:151–60.
[9] Hermann BP, Jones JE, Sheth R, Koehn M, Becker T, Fine J, et al. Growing up with
epilepsy: a two-year investigation of cognitive development in children with new
onset epilepsy. Epilepsia 2008;49:1847–58.
[10] Lin JJ, Mula M, Hermann BP. Uncovering the neurobehavioural comorbidities of
epilepsy over the lifespan. Lancet 2012;380:1180–92.
[11] Baca CB, Vickrey BG, Caplan R, Vassar SD, Berg AT. Psychiatric and medical comor-
bidity and quality of life outcomes in childhood-onset epilepsy. Pediatrics 2011;
128:e1532-3.
[12] Linehan C, Tellez-Zenteno JF, Burneo JG, Berg AT. Future directions for epidemiology
in epilepsy. Epilepsy Behav 2011;22:112–7.
 A. Oh et al. / Epilepsy & Behavior xxx (2017) xxx–xxx
[13] Chin RF, Cumberland PM, Pujar SS, Peckham C, Ross EM, Scott RC. Outcomes of child-
hood epilepsy at age 33 years: a population-based birth-cohort study. Epilepsia
2011;52:1513–21.
[14] Jones JE, Watson R, Sheth R, Caplan R, Koehn M, Seidenberg M, et al. Psychiatric
comorbidity in children with new onset epilepsy. Dev Med Child Neurol 2007;49:
493–7.
[15] Berg AT, Smith SN, Frobish D, Levy SR, Testa FM, Beckerman B, et al. Special educa-
tion needs of children with newly diagnosed epilepsy. Dev Med Child Neurol 2005;
47:749–53.
[16] Hermann B, Jones J, Dabbs K, Allen CA, Sheth R, Fine J, et al. The frequency, compli-
cations and aetiology of ADHD in new onset paediatric epilepsy. Brain 2007;130:
3135–48.
[17] Thurman DJ, Beghi E, Begley CE, Berg AT, Buchhalter JR, Ding D, et al. Standards for
epidemiologic studies and surveillance of epilepsy. Epilepsia 2011;52(Suppl. 7):
2–26.
[18] Holden EW, Grossman E, Nguyen HT, Gunter MJ, Grebosky B, Von Worley A, et al.
Developing a computer algorithm to identify epilepsy cases in managed care organi-
zations. Dis Manag 2005;8:1–14.
[19] Helmers SL, Thurman DJ, Durgin TL, Pai AK, Faught E. Descriptive epidemiology of
epilepsy in the U.S. population: a different approach. Epilepsia 2015;56:942–8.
[20] Kim H, Thurman DJ, Durgin T, Faught E, Helmers S. Estimating epilepsy incidence
and prevalence in the US pediatric population using nationwide health insurance
claims data. J Child Neurol 2016;31:743–9.
Please cite this article as: Oh A, et al, Comorbidities and risk factors associated with newly diagnosed epilepsy in the U.S. pediatric population,
Epilepsy Behav (2017), http://dx.doi.org/10.1016/j.yebeh.2017.07.040
7
[21] United States Census Bureau. Metropolitan and micropolitan: 2010 and 2000 popu-
lation tables. http://www.census.gov/population/metro/data/index/html/; 2012.
[accessed 16.10.01].
[22] Aaberg KM, Bakken IJ, Lossius MI, Lund Soraas C, Haberg SE, Stoltenberg C, et al.
Comorbidity and childhood epilepsy: a nationwide registry study. Pediatrics 2016;
138. http://dx.doi.org/10.1542/peds.2016-0921.
[23] Wagner JL, Wilson DA, Smith G, Malek A, Selassie AW. Neurodevelopmental
and mental health comorbidities in children and adolescents with epilepsy and
migraine: a response to identified research gaps. Dev Med Child Neurol 2015;57:
45–52.
[24] Whitehead E, Dodds L, Joseph KS, Gordon KE, Wood E, Allen AC, et al. Relation of
pregnancy and neonatal factors to subsequent development of childhood epilepsy:
a population-based cohort study. Pediatrics 2006;117:1298–306.
[25] Kulak W, Sobaniec W. Risk factors and prognosis of epilepsy in children with cere-
bral palsy in north-eastern Poland. Brain Dev 2003;25:499–506.
[26] Zelnik N, Konopnicki M, Bennett-Back O, Castel-Deutsch T, Tirosh E. Risk factors for
epilepsy in children with cerebral palsy. Eur J Paediatr Neurol 2010;14:67–72.
[27] KFF (Kaiser Family Foundation). Health insurance coverage of children 0–18.
http://kff.org/state-category/health-coverage-uninsured/health-insurance-status/;
2013. [accessed 16.11.15].